2025.3.11 シグモイド関数

$ \sigma(x) = \frac{1}{1+\exp(-x)}

実装したもの：

code:p1.py

import numpy as np

import matplotlib.pyplot as plt

def sigmoid(x):

return 1.0 / (1.0 + np.exp(-x))

x = np.linspace(-10, 10)

y = sigmoid(x)

plt.grid()

plt.plot(x, y)

plt.show()

PyTorch

code:p1.py

import matplotlib.pyplot as plt

import torch as pt

x = pt.linspace(-10, 10, 50) # 第三引数省略不可

y = pt.sigmoid(x) # 第１引数はtensor型

plt.grid()

plt.plot(x, y)

plt.show()

簡単なメモ

$ f_1 = -x

$ f_2= \exp(-x)

$ f_3 = 1+\exp(-x)

$ f_4 = \frac{1}{1+\exp(-x)}

code:p1.py

import matplotlib.pyplot as plt

import numpy as np

def f1(x):

return -x

def f2(x):

return np.exp(-x)

def f3(x):

return 1 + np.exp(-x)

def f4(x):

return 1/(1 + np.exp(-x))

x = np.linspace(-5, 5)

plt.plot(x, f1(x), label='$-x$')

plt.plot(x, f2(x), label='$\exp(-x)$')

plt.plot(x, f3(x), label='$1+\exp(-x)$')

plt.plot(x, f4(x), label='$1/(1+\exp(-x))$')

plt.ylim(-0.5, 2)

plt.grid()

plt.legend()

plt.show()

https://scrapbox.io/files/67cfe2f9a14ce0bc91ca3449.png